Super Converged Infrastructure has been accepted by more and more users, and is considered to be the future development trend of IT infrastructure. Many users are already considering how to smoothly migrate to a production environment over converged infrastructure, and how to build ultra-integration of VMware vSphere infrastructure around the existing ecology. Although the super integration architecture should support a variety of Hypervisor, but companies still tend to build over converged infrastructure around VMware vSphere, for two main reasons:
Compatibility with VMware vSphere's business system had already been fully validated.
VMware vSphere has trained a group of professional operation and maintenance personnel, have adequate technical reserves to worry about replacing the virtualization platform will have to pay a larger study costs.
Eco-around VMware vSphere already have a lot of input, such as: VMware vSphere licensing, backup software, monitoring software, and even cloud platform integration has been completed and so on.
So what precautions selection of ultra-converged infrastructure build around VMware vSphere need it? Since the Hypervisor has been specified, then the key is the selection of a distributed storage software, so the need to meet at least the following factors:
Better compatibility with VMware vSphere mainstream version, ensure that the entire platform stability, and have the ability to continue to develop.
Specifically optimized for VMware vSphere operating environment, provide superior I / O performance.
Online scale-out capabilities, processes need to be simple enough.
Because it is ultra deployment of converged, distributed storage software required to ensure the normal operation of the premise footprint as low as possible, so you can ensure that more resources available to a virtual machine.
Provide storage support for advanced features, such as storage synchronous, asynchronous replication and other advanced features to meet the production requirements of business continuity.
SMTX distributed block storage (ZBS) based on vSphere to build ultra-converged infrastructure
SMTX distributed block storage Introduction
SMTX distributed block storage module (internal code "ZBS", hereinafter referred SmartX The ZBS) is SmartX self-developed high-performance, highly reliable distributed mass storage system. SmartX ZBS self is born from the fusion of architecture for ultra design. Full advantage of the virtualization platform features and characteristics of SSD devices, in-depth optimization and integration. Support and KVM, VMware ESXi, XenServer and other virtualization / cloud integration platform combine to build super platform.
SmartX ZBS has the following functions and features:
high performance
In SSD devices as the cache, that take advantage of low latency SSD, the high IOPS;
Support ultra converged infrastructure, computing and storage running on the same physical server, storage, computing perception, most of the I / O can be done locally to avoid network latency.
High reliability
By multiple copies, ensure data reliability in server downtime and business continuity;
Write multiple copies using a strong consistency to ensure data consistency.
Low consumption
Commodity x86-compatible server, no special hardware requirements;
Storage only with limited computing resources (6 vCPU 16GB memory), to avoid the additional purchase of a storage server, a perfect fit over converged infrastructure.
Laterally expandable
Synchronous linear expansion performance and capacity;
The largest single cluster can support up to 255 nodes.
Architecture
SmartX based on VMware vSphere deployment of ultra-integration, the main way is to deploy VMware ESXi on the bare metal servers, and SCVM (SmartX Controller Virtual Machine) is deployed on VMware ESXi, SCVM the name suggests is to act as a virtual machine storage controller role, it by passthrough technology (to avoid possible through the ESXi virtualization performance loss) directly to hard disk resource management server (including SSD and HDD).
Between multiple SCVM to Gigabit networking built into a unified storage pool to consolidate storage resources with built-in SmartX ZBS distributed storage software, and then use the NFS protocol to mount a VMware ESXi datastore provides storage services.
SMTX distributed block storage (ZBS) and combine the advantages of customer value vSphere
1. Good hardware compatibility allowing users to choose more flexible and stable
ZBS fully operational at the application layer, directly call native CentOS drive is compatible with most of the hardware available in the market, can effectively alleviate vSAN need to maintain a separate hard drive, it can not be unified with vSphere HCL selection problems and stability to the user's risk.
2. Hypervisor version and users loosely coupled reduce upgrade costs
Adopted under ZBS over converged infrastructure in VMware SCVM deployment mode, ZBS and vSphere version of the adapter more flexible and reduce the extra cost of native programs vSAN need to upgrade vSphere synchronization upgrade brings.
3. Data Localization bring better performance and service experience
ZBS In addition to supporting SSD cache acceleration, but also for VMware vSphere virtual environment optimized replica placement strategy, ZBS by storing redundant copies of data protection technology VM, and data localization function which will be a full copy of the VM is running node access, another copy is stored in a remote node to ensure data redundancy. The advantage of this node is a VM to access data shortens the I / O path, reduce latency and reduce the pressure on network bandwidth, bring substantial lifting I / O performance.
Meanwhile the case, ZBS perfect support for VMware vMotion functionality, even if the VM live migration occurs, the data localization function is still valid, the system will automatically detect the VM migration occurred because the remote I / O access, and not a temporary migration occurs automatically data migration, re-formation of localized data in the new node.
4. Complete high availability mechanisms to ensure the stable operation of the user service
ZBS support VMware HA functionality in response to node failure scenarios, I / O routing redirection for the failure of other hardware, providing finer than HA high availability features.
(1) failure to deal with the storage network
When storage network failure, SmartX Halo support storage networking links automatically redirected to the service network, acts as a temporary storage link data transmission, and will not trigger VMware HA function, a virtual machine on a host of business and will not happen interrupted. The same situation, VMware vSAN storage network failure directly trigger VMware HA functionality to virtual machines on that node forced to close and move to the other node to restart the virtual machine, so business disruptions and migrations happening.
(2) SCVM failure to respond
ZBS and VMware ESXi are loosely coupled relationship, which makes the HA mechanism more flexible, even SCVM failure can not provide storage services, while under ESXi itself is not the fault of the situation, SmartX trigger I / O redirection, making the node virtual machines can be done by a virtual machine I / O requests quickly switch to SCVM on the remote node to finish, so you can do the same in the case of virtual machines without service interruption, and to continue to provide storage services.
5. Rapid Cloning effectively enhance the efficiency of the mechanism
ZBS support snapshot replication and cloning technology write COW, when performing the snapshot and clone operations, copying real data does not occur only when the data modify data replication will occur, so the speed is very fast execution clone, you can cloning completed within a few seconds, and the time is not VM clone size affects memory capacity. Meanwhile, the VAAI ZBS VMware support, may be cloned and VMware snapshot operations to offload ZBS underlying execution, so that execution vCenter rapid cloning techniques are equally applicable VM cloning of ZBS.
6. disaster recovery and business continuity feature provides greater reliability of data
ZBS support deployment across the room Replication Clusters based on VMware vSphere, the environment can be achieved with the application of "double live." Also support the underlying storage asynchronous replication, the host may be switched when the room to spare room when the fault recovery operations.
SmartX visit the official website to learn more deployment solution based on converged architecture super VMware virtualization platform: https://www.smartx.com/solution